Species Identification and Antifungal Susceptibility of Candida Bloodstream Isolates from 1 ...

31 Pages·2012·0.49 MB·English

JCM Accepts, published online ahead of print on 8 August 2012
J. Clin. Microbiol. doi:10.1128/JCM.01283-12
Copyright © 2012, American Society for Microbiology. All Rights Reserved.
1 Species Identification and Antifungal Susceptibility of Candida Bloodstream Isolates from
2 Population-based Surveillance in Two US Cities: 2008-2011
3
4
5
6
7
8
D
9 o
w
10
n
11 lo
12 Shawn R. Lockhart1, Naureen Iqbal1, Angela M. Ahlquist1, Monica M. Farley2,3, Lee H. a
d
13 Harrison4, Carol B. Bolden1,Wendy Baughman2 Betsy Stein3, Rosemary Hollick4, Benjamin J. e
d
14 Park1, and Tom Chiller1
f
r
15 o
16 1Mycotic Diseases Branch, Centers for Disease Control and Prevention, Atlanta, GA, 2Atlanta m

17 Veterans Affairs Medical Center, Atlanta, GA, 3Emory University, Atlanta, GA, 4Johns Hopkins h
t
t
18 Bloomberg School of Public Health, Baltimore MD p
:
19 //
20 jcm
21 .a
s
22 m
23 .o
r
24 g
/

25 o
n
26 D
27 ec
e
28 m
29 b
e
30 r
2
31 7
,
32 *Send all correspondence to: 2
0
33 Dr. Shawn R. Lockhart 1
34 Mycotic Diseases Branch 8

b
35 Centers for Disease Control and Prevention y

36 1600 Clifton Rd. g
u
37 Mailstop G-11 e
38 Atlanta, GA 30333 s
t
39 Phone: [phone]
40 FAX: [phone]
41 E-mail: [email protected]
42
43
44
45 ABSTRACT
46 Between 2008 and 2011, population-based candidemia surveillance was conducted in Atlanta,
47 GA and Baltimore, MD. Surveillance had been previously performed in Atlanta in 1992-1993
48 and in Baltimore in 1998-2000, making this the first population-based candidemia surveillance
D
49 conducted over multiple time points in the US. From 2,675 identified cases of candidemia in the o
w
n
50 current surveillance, 2,329 Candida isolates were collected. Candida albicans no longer lo
a
d
51 comprised the majority of isolates but remained the most frequently isolated species (38%), e
d

f
r
52 followed by C. glabrata (29%), C. parapsilosis (17%) and C. tropicalis (10%). The species o
m

53 distribution has changed over time; in both Atlanta and Baltimore the proportion of C. albicans h
t
t
p
:
54 decreased and the proportion of C. glabrata increased, while the proportion of C. parapsilosis //
jc
m
55 increased in Baltimore only. There were 98 multi-species episodes, with C. albicans and C. .
a
s
m
56 glabrata the most frequently encountered combination. The new species-specific CLSI Candida .
o
r
g
57 MIC breakpoints were applied to these data. With the exception of C. glabrata (11.9% resistant), /

o
n
58 resistance to fluconazole was very low (2.3% for C. albicans, 6.2% for C.tropicalis and 4.1% for D
e
c
59 C. parapsilosis). There was no change in the proportion of fluconazole resistance between e
m
b
60 surveillance periods. Overall echinocandin resistance was low (1%) but was higher for C. e
r

2
7
61 glabrata isolates, ranging from 2.1% resistant to caspofungin in Baltimore to 3.1% resistant to
,

2
0
62 anidulafungin in Atlanta. Given the increase at both sites and the higher echinocandin resistance, 1
8

b
63 C. glabrata should be closely monitored in future surveillance. y

g
u
64 e
s
t
65
66
67
68 INTRODUCTION
69 Candida species are a leading cause of healthcare-associated bloodstream infections
70 (BSIs) in the U.S., accounting for approximately 10% of healthcare-associated BSIs and one of
71 the highest crude mortality rates (37). Candidemia infections increase the risk of patient
D
72 mortality and increase both the length of stay and cost associated with hospitalization (11,17). o
w
n
73 To implement appropriate control measures, it is important to understand the epidemiology of lo
a
d
74 BSIs caused by Candida species. Although the species distribution has changed over the last e
d

f
r
75 three decades, more than 95% of candidemia infections have been caused by the five most o
m

76 common species: Candida albicans, C. glabrata, C. parapsilosis, C. tropicalis and C. krusei h
t
t
p
:
77 (4,26,36). The most dramatic temporal change has been the shift from C. albicans to non-C. //
jc
m
78 albicans species as causative agents of the majority of BSIs (1,4,10,12,16,26,36). The .
a
s
m
79 significance of this trend is that non-C. albicans species, especially C. glabrata, can be less .
o
r
g
80 susceptible to fluconazole, the most inexpensive and readily available antifungal agent used to /

o
n
81 treat candidemia (10,25,28,31). The shift to non-C. albicans species is also significant because D
e
c
82 the newest class of antifungal agents, the echinocandins, are not currently recommended as e
m
b
83 primary therapy against C. parapsilosis, the third most common species in the US (10,16,20,25). e
r

2
7
84 Sentinel and population-based surveillance play complementary roles in our
,

2
0
85 understanding of candidemia. Sentinel surveillance can encompass many facilities over broad 1
8

b
86 geographic areas but the makeup of individual institutions may not be representative of whole y

g
u
87 populations. The advantage of population-based surveillance is that it allows the calculation of e
s
t
88 population-based incidence rates (9,10,12,16,25,28,29). Antifungal susceptibility testing of
89 Candida isolates collected as part of surveillance has played an increasingly important role in
90 surveillance strategies for epidemiological evaluation of antifungal resistance
91 (2,6,9,10,12,28,33,34). With the advent of standardized susceptibility testing methodology (7),
92 past and present surveillance can be directly compared to monitor temporal trends in antifungal
93 resistance. Global sentinel surveillance revealed an increasing trend in resistance to fluconazole
94 among C. parapsilosis, C. guilliermondii and C. lusitaniae, while the resistance rate among C.
D
95 albicans, C. glabrata and C. tropicalis fluctuated slightly or remained stable (28). Because of o
w
n
96 the recent appearance of echinocandins in clinical practice, temporal surveillance data for lo
a
d
97 susceptibility to these drugs are limited, but an early report indicated that susceptibility rates e
d

f
r
98 were stable (23). Temporal changes in antifungal susceptibility for Candida species to any o
m

99 antifungal agent have not yet been described using population-based surveillance in a single US h
t
t
p
:
100 city. //
jc
m
101 The Centers for Disease Control and Prevention (CDC) and state partners in the .
a
s
m
102 Emerging Infections Program (EIP) have undertaken two active, population-based surveillance .
o
r
g
103 studies to determine the incidence of candidemia, the distribution of Candida species causing /

o
n
104 BSI, and the prevalence of antifungal drug resistance (10,12). In both studies, surveillance was D
e
c
105 conducted in two geographic areas: Atlanta, GA (1992-1993); and Baltimore city and county, e
m
b
106 MD (1998-2000). CDC and EIP partners again conducted population-based surveillance for e
r

2
7
107 candidemia in 2008-2011, returning to the previous areas of Atlanta, GA and Baltimore, MD. By
,

2
0
108 returning to previous areas of surveillance, changes in species distribution and antifungal 1
8

b
109 susceptibility can be described. Here we report the laboratory results of that surveillance y

g
u
110 including molecular determination of Candida species and antifungal susceptibility testing e
s
t
111 against nine antifungal agents. This is the first time that susceptibility to echinocandins has been
112 determined within a US population-based collection of isolates.
113
114 MATERIALS AND METHODS
115 Case and isolate definitions. Isolates were obtained from persons with an incident episode of
116 candidemia (defined below) identified between March 1, 2008 and February 28, 2011 for
117 residents of Georgia Health District 3 (the eight counties comprising metropolitan Atlanta,
D
118 GA)(population=3.8 million) and between June 1, 2008 and May 31, 2011 for residents of o
w
n
119 Baltimore City or Baltimore County, Maryland (population=1.4 million). The isolate retrieval lo
a
d
120 rate from identified incident cases in Atlanta was 71% and in Baltimore it was 92%. An e
d

f
r
121 incident episode of candidemia was defined as the 30 days following the first positive blood o
m

122 culture for Candida species in a resident of the surveillance area. Positive cultures of the same h
t
t
p
:
123 species within the 30-day period were considered part of the incident case episode and were not //
jc
m
124 captured but positive cultures of a different species during a single episode were captured. .
a
s
m
125 Cultures drawn more than 30 days after the incident case were considered a new incident episode .
o
r
g
126 and assigned a new case number. Isolates were received from 17 hospitals in Baltimore with an /

o
n
127 average of 87 isolates per hospital (range of 1-562) and from 24 hospitals in Atlanta with an D
e
c
128 average of 63 isolates per hospital (range of 2-310). e
m
b
129 Isolate storage, DNA extraction, PCR amplification and sequencing. Prior to use, all isolates e
r

2
7
130 were stored in glycerol at -70° C. Isolates were identified by both conventional biochemical
,

2
0
131 means at the referring institutions and molecular means using a Luminex assay or DNA 1
8

b
132 sequencing of the D1/D2 subunit of the 28S rDNA at the CDC as previously described (8). y

g
u
133 Antifungal susceptibility testing. Antifungal susceptibility testing was performed by broth e
s
t
134 microdilution with fluconazole, itraconazole, voriconazole, posaconazole, flucytosine,
135 anidulafungin, caspofungin, and micafungin as described by the Clinical and Laboratory
136 Standards Institute (CLSI) M27-A3 document guidelines (7) using frozen RPMI microbroth
137 trays custom manufactured by TREK Diagnostics (Cleveland, OH). Susceptibility testing for
138 amphotericin B was performed by Etest as per the manufacturer’s instructions (bioMérieux,
139 Durham, NC). Results were read visually after 24 h for fluconazole, voriconazole, posaconazole
140 (24), caspofungin, micafungin, anidulafungin, amphotericin B and flucytosine (14) and 48 h for
D
141 itraconazole. For all but amphotericin B, the MIC was read as the lowest concentration of drug o
w
n
142 that caused a significant decrease in growth as compared to the control well. Recently approved lo
a
d
143 CLSI 24 h resistance breakpoints for fluconazole, voriconazole and the echinocandins were used e
d

f
r
144 (21,22,27): isolates of C. albicans, C, tropicalis and C. parapsilosis with an MIC ≥8 µg/ml and o
m

145 isolates of C. glabrata with an MIC ≥64 µg/ml were considered resistant to fluconazole; C. h
t
t
p
:
146 krusei was considered intrinsically resistant to fluconazole; isolates of C. albicans, C, tropicalis //
jc
m
147 and C. parapsilosis with an MIC ≥1 µg/ml and C. krusei isolates with an MIC ≥2 µg/ml were .
a
s
m
148 considered resistant to voriconazole; isolates of C. albicans, C. tropicalis and C. krusei with an .
o
r
g
149 MIC ≥1 µg/ml were considered resistant to caspofungin, micafungin and anidulafungin; isolates /

o
n
150 of C. parapsilosis with an MIC ≥8 µg/ml were considered resistant to caspofungin, micafungin D
e
c
151 and anidulafungin; isolates of C. glabrata with an MIC ≥0.5 µg/ml were considered resistant to e
m
b
152 caspofungin and anidulafungin, while those with an MIC ≥0.25 µg/ml were considered resistant e
r

2
7
153 to micafungin. For C. dubliniensis, the C. albicans breakpoints were used. All other species
,

2
0
154 were considered to have no current breakpoints. Amphotericin B susceptibility was performed 1
8

b
155 using Etest as per the manufacturer’s instructions and MIC values were read at 24 h. Quality y

g
u
156 control isolates C. parapsilosis ATCC 22019 and C. krusei ATCC 6258 were included on each e
s
t
157 day of testing. As prior population-based surveillance was conducted in both Atlanta (1992-
158 1993)(12) and Baltimore (1998-2000)(10) and susceptibility to fluconazole was previously
159 determined for all isolates, we applied the new breakpoints to the previous surveillance results
160 and compared them to the current results.
161 Multi-species episodes. A multispecies episode was defined as when a Candida species other
162 than the incident isolate species was recovered from a patient within the 30 days encompassing
D
163 the incident episodic period or when two species were recovered simultaneously as part of the o
w
n
164 incident episode. lo
a
d
165 Defining multidrug resistance. To screen for multidrug-resistant (MDR) isolates, the most e
d

f
r
166 frequently used antifungal agents in our surveillance areas were selected; fluconazole and o
m

167 voriconazole representing the azoles, caspofungin, micafungin and anidulafungin representing h
t
t
p
:
168 the echinocandins and amphotericin B as the polyene. Isolates were considered multidrug //
jc
m
169 resistant if they were resistant to one or more representative antifungal(s) of two of the above .
a
s
m
170 classes. Since there are no resistance breakpoints for amphotericin B, the arbitrary but accepted .
o
r
g
171 breakpoint of 2.0 μg/ml for resistance was used. /

o
n

172 D
e
c
173 RESULTS e
m
b
174 Cases. A total of 2,227 incident episodes had one or more isolates available, from which 2,329 e
r

2
7
175 eligible isolates were collected (Table 1). Incident isolates were available for 2,136 (80%) ,

2
0
176 patients. There were 94 episodes with two species, 42 in Atlanta and 52 in Baltimore, one 1
8

b
177 episode with three species, one episode with four species, and one episode with five species, all y

g
u
178 in Baltimore. Candida albicans was the most frequently isolated species, comprising 34% of the e
s
t
179 isolates from Baltimore and 41% of the isolates from Atlanta (Figure 1). Candida glabrata was
180 the second most prevalent species from both sites (27% in Atlanta and 31% in Baltimore)
181 followed by C. parapsilosis and C. tropicalis. In Baltimore, C. dubliniensis was the fifth most
182 prevalent species but in Atlanta, C. lusitaniae was the fifth most prevalent species. Overall, C.
183 albicans, C, glabrata, C. parapsilosis, C. tropicalis and C. krusei comprised 95% of the isolates;
184 with the addition of C. dubliniensis and C. lusitaniae, they comprised 98% of the isolates. Two
185 hundred seventeen (8%) case-patients had >1 30 day episode of candidemia. Of those, 69 (32%)
D
186 were C. glabrata, 45 (21%) were C. albicans, 43 (20%) were C. parapsilosis, and 29 (13%) were o
w
n
187 C. tropicalis.” lo
a
d
188 e
d

f
r
189 Susceptibility. Among the 2329 isolates tested (Tables 2,3), overall resistance to fluconazole o
m

190 was 7.3%, primarily due to C. glabrata and C. krusei, which together comprised 31% of the total h
t
t
p
:
191 isolates for which breakpoints were available but 68% of the fluconazole resistant isolates. //
jc
m
192 Resistance to fluconazole varied by species with C. glabrata having the highest rate (11.9%) .
a
s
m
193 followed by C. tropicalis (6.2%), C. dubliniensis (5.6%), C. parapsilosis (4.1%), and C. albicans .
o
r
g
194 (2.3%). Resistance to itraconazole was high (21.1%), again primarily due to C. glabrata (59.6% /

o
n
195 of which were resistant) and C. krusei (21.9% resistant). Overall voriconazole resistance was D
e
c
196 low (1.0%), ranging from 0% among C. krusei to 2.5% among C. tropicalis isolates. There are e
m
b
197 no currently approved breakpoints for posaconazole but the MIC value for most species was e
90 r

2
7
198 ≤0.5 µg/ml (Table 2).
,

2
0
199 Resistance to echinocandins was low overall, ranging from 1.0% to caspofungin or 1
8

b
200 anidulafungin to 1.1% for micafungin. Again, resistance was primarily found in C. glabrata y

g
u
201 which accounted for 65% (20/31) of the isolates resistant to any echinocandin, followed by C. e
s
t
202 albicans (19% of the resistant isolates).
203 There are no established breakpoints for amphotericin B but an MIC breakpoint of 2
204 µg/ml is widely considered to be elevated. Only four isolates had amphotericin B MIC values of
205 2 µg/ml: two C. glabrata isolates and one isolate each of C. tropicalis and C. parapsilosis.
206
D
207 Susceptibility by surveillance site. The overall proportion of isolates that were resistant to o
w
n
208 fluconazole was slightly higher in Atlanta than it was in Baltimore. When analyzed by species, lo
a
d
209 fluconazole resistance was nearly identical for C. albicans in Baltimore and Atlanta (Tables 4A e
d

f
r
210 and 4B). Fluconazole resistance was higher in Atlanta among C. glabrata, C. tropicalis and C. o
m

211 parapsilosis, and nearly twice as high in Atlanta as in Baltimore for the latter two species. h
t
t
p
:
212 Overall resistance to voriconazole was again almost identical in Atlanta and Baltimore and only //
jc
m
213 C. tropicalis isolates in Atlanta (3.8%) had resistance higher than 2%. Overall resistance to .
a
s
m
214 echinocandins was the same at the two sites (Tables 4A and 4B). .
o
r
g
215 Temporal comparison of susceptibility results. We compared the previously published /

o
n
216 population-based surveillance results (10,12) to the current results (Table 5). Resistance to D
e
c
217 fluconazole in C. albicans remained relatively unchanged in both Atlanta and Baltimore. e
m
b
218 Fluconazole resistance among C. glabrata isolates in Atlanta was 20% in 1993 and 13.2% in e
r

2
7
219 2011, however there were only 35 isolates available for testing in 1993, just over 10% of the
,

2
0
220 number of isolates available for testing in 2011. Fluconazole resistance among C. glabrata 1
8

b
221 isolates in Baltimore was 8.0% in 2008 and 10.8% in 2011. Resistance of C. parapsilosis to y

g
u
222 fluconazole was only slightly different over time in both Atlanta (from 1.9% to 5.6%) and e
s
t
223 Baltimore (from 0% to 2.3%) and the same was true for C. tropicalis in Atlanta (from 3.3% to
224 8.6%) and in Baltimore (from 8.8% to 4.3%).
225 Multidrug-resistant isolates. Among isolates with acquired resistance, there were three MDR
226 isolates from Atlanta, all C. glabrata resistant to fluconazole and one or more echinocandin.
227 There were six MDR isolates From Baltimore, five C. glabrata and one C. albicans all resistant
228 to fluconazole and one or more echinocandin.
D
229 o
w
n
230 Multi-species incident episodes. There were 42 multi-species episodes in Atlanta and 55 in lo
a
d
231 Baltimore. In Atlanta, there were 23 episodes where two species were isolated simultaneously, e
d

f
r
232 and 19 episodes where two species were isolated on different days during the 30 day episodic o
m

233 period (range 1-27 days; average of 9 days apart). In Baltimore, there were 36 episodes with two h
t
t
p
:
234 species isolated simultaneously (including one patient with three species simultaneously) and 21 //
jc
m
235 episodes where two species were isolated on different days (including one patient with four .
a
s
m
236 species isolated on four different days and one patient with two simultaneous species, then a .
o
r
g
237 different species three days later and then two different species again simultaneously) during the /

o
n
238 30 day episodic period (range 1-26 days; average of 7 days apart). Candida albicans was the D
e
c
239 most common species occurring in multispecies episodes, in Baltimore most often with C. e
m
b
240 glabrata (27 times) followed by C. parapsilosis (six times) and in Atlanta most often with C. e
r

2
7
241 glabrata (12 times) or C. parapsilosis (seven times). Candida glabrata occurred with C.
,

2
0
242 parapsilosis six times in Atlanta but only once in Baltimore. 1
8

b
243 y

g
u
244 DISCUSSION e
s
t
245 There were a number of important results from this population-based surveillance. First,
246 the distribution of species at both of the surveillance sites has changed, with the proportion of C.
247 albicans infections decreasing and the proportion of C. glabrata and C. parapsilosis infections